Precipitation of Chinese script phase

Earlier work [1] on the effect of cooling rate on the morphology of beta phase from Al-6.5Si-1Fe showed an anomaly in the precipitation of iron rich intermetallic phase.

Metallographic examination of DTA sample with cooling rate of 0.2°C/min shows the performed by recording the DTA signal with various low scanning rates for cooling of 0.02, 0.05 and 0.1°C/min from 620 to 590°C. The low scanning rate was applied from above the liquidus until beta phase precipitation was deemed to be finished based on previous experiments. After that, the DTA was cooled at 2°C/min. The sample was held at 620°C for 30 min before the cooling sequence was started.

Figure 6.1. DTA thermograms recorded at very low cooling rates (the graphs are adjusted along the y axis to separate the curves with the lowest scanning rate at the bottom).

DTA thermograms in Figure 6.1 illustrate distinctive thermal arrests at the start of the cooling process which relate to the nucleation and growth of primary (Al) dendrites (blue arrows) and the beta phase precipitation reaction L→ (Al) + β (red arrows). In between those two major peaks, a small peak (black arrow) was also detected in the sample with 0.05°C/min cooling rate which could be thought to be associated to the precipitation of Chinese script phase. However, that peak was not seen in other samples.

Moreover, the thermograms show a noise increase after the primary (Al) arrest. It is further seen that the noise amplitude is increased as the cooling rate increases. The likely explanation for this is related to the DTA principle which is based on measuring the signal difference between the sample and a reference being given the same heating or cooling rate.

The change in thermal resistance at metal and crucible interface could cause this noise differences. Nevertheless, there also other possibility such as thermocouple characteristic in the DTA which used electric potential difference that could have been causing the noise.

-4 -2 0 2 4

590 595 600 605 610 615 620

Temperature (°C)

6.1.2. Characterization: EDS analysis, EPMA, EBSD andXRD study

The DTA samples were prepared for metallographic observation under optical microscope and SEM. Micrographs in Figure 6.2 and 6.3 clearly show the presence of Chinese script phase, plate-like beta phase and eutectic silicon.

a.

b.

c.

Figure 6.2. Micrographs of DTA samples for various cooling rates showing the presence of script phase in interdendritic regions (a) 0.02°C/min, (b) 0.05°C/min, and (c) 0.1°C/min (images to the right showed

enlarged red circled areas).

The phase morphology which is characterized with fishbone or script shape appears in interdendritic regions, either in the middle or near to the surface of the sample. The size of the script precipitates varies but overall was less than 100 m in length in 2D section. From the metallographic examination we could also notice that no or very few plate-like beta phase could be detected near script precipitates.

Figure 6.3. SEM micrographs showing clusters of script phase, taken from DTA sample with various cooling rate (a) 0.02°C/min (b) 0.05°C/min and (c) 0.1°C/min.

6.1.2.1 EDS and micro-probe analysis

EDS analysis of the chinese script phase was conducted on various locations as illustrated in Figure 6.4. The result revealed the presence of Al, Si and Fe and no minor element detected. The EDS quantitative analysis showed that script phase has a similar chemical composition to beta phase, as seen in Table 6.1.

Figure 6.4 . SEM images showing script phase precipitates and the measurement locations taken from sample with cooling rate of 0.1°C/min.

Table 6.1. EDS results of script phase.

Test location No.

Chemical composition (wt%)

Al Si Fe

3 56.4 14.7 24.8

4 55.8 14.9 25

8 55.7 14.4 25

9 57.8 14.7 24.7

Average 56.4 14.7 24.9

Further analysis with electron micro probe analyzer (EPMA) was performed on the same sample and some locations corresponded to the previous EDS measurement, as seen in Figure 6.5. EPMA results as seen in Table 6.2 do not significantly differ from EDS results.

The analysis reveals that the Si and Fe content in Chinese script phase was quite high with Fe/Si ratio around 1.8, thus the script phase could be classified into beta phase.

Figure 6.5. EPMA test point location of two script phase from the same sample as in Figure 4.

Table 6.2. EPMA result of script phase.

Phases Test

Analysis with EDS on the Chinese script phase also revealed the presence of cerium rich precipitates which appear in bright contrast in SEM as seen in the example of Figure 6.6.

2 1

6.1.2.2. Micro XRD

X-ray analysis was performed using a micro XRD method (Bruker Advance D8 equipped with micro beam focus) on the area as seen in Figure 6.7. This method was used to confirm the previous metallographic result by crystal structure identification relating to JCPDS file. The XRD used a Cu-Kα radiation (λ = 1.54060 Å) with source parameters of 40 mA and 40 kV. Diffraction pattern was acquired between 15 and 80° in 2 theta with a step size of 0.03^oand exposure area (slit size) of 100x100 m². Due to the use of micro-beam, the X-ray flux generated was very small, therefore the peak acquired was not quite high and required a longer and repeated acquisition to ensure sufficient recorded data.

Figure 6.7. Location of the micro XRD beam.

Four major peaks were observed between 25 and 60° in 2 theta as seen in Figure 6.8.

Based on a structural database search, those peaks appear to correspond to Al (111) and Si (111), (220), (311) as indexed with JCPDS cards no. 00-004-0787 and 00-027-1402 respectively. Very small additional peaks could be detected which are marked with solid symbols in Figure 6.8. The peaks could be indexed with JCPDS cards file no. 01-071-0238 which corresponds to Al167.8Fe44.9Si23.9 phase and JCPDS card 00-054-0376 which corresponds to Al9Fe2Si2. It was also verified that these peaks could not be indexed as alpha phase.

Figure 6.8. XRD patterns showing peaks of (Al), Si and script phase

6.1.2.3. EBSD analysis

In order to confirm the XRD result, further examination with EBSD was performed on Chinese script precipitates. Example of un-indexed and indexed Kikuchi pattern of the script phase can be seen in Figure 6.9. Its analysis indicates the script phase is monoclinic as -Al9Fe2Si2 while indexing could not match for hexagonal -Al8Fe2Si. Median angular deviation (MAD) index method gives 0.76 with 10-band detected match for beta phase monoclinic structure.

0 200 400 600 800 1000 1200 1400

20 30 40 50 60 70 80

2 theta

intensity

 : Al

 : Si

 : script phase

a.

b. c.

Figure 6.9. EBSD pattern obtained from a script precipitates. (a) micrograph (b) unindexed pattern, (c) indexed pattern. The EBSD pattern is indexed according to the monoclinic β-Al9Fe2Si2.

6.1.3. Discussion

Based on the experiment results, we would like to assess the formation of script phase by considering the solidification path and the effect of element to their formation. The solidification path of Al-6.5Si-1Fe was calculated assuming the solidification process follows lever rule or Scheil model as seen in previous chapter. Based on the solidification path, the only intermetallic phase appearing during solidification should be beta. However, as seen from the metallographic analysis, some Chinese script precipitates which are commonly associated to alpha phase with hexagonal (Al₈Fe₂Si) or cubic (Al₁₅(Fe,Mn)₂Si₂) structure.

According to the Al-Fe-Si ternary phase diagram one would expect any precipitate of hexagonal alpha should transform to monoclinic beta by a peritectic reaction during solidification. Such a transformation is illustrated in Figure 6.10 from Gorny et al. [2] in the case of sample cooled at 6°C/min. However, metallographic observation by SEM in the present work did not show any indication of phase transformation by peritectic reaction.

25m

Figure 6.10. Peritectic reaction of alpha τ₅to beta τ₆phase, taken from Gorny et al. [2].

Though no sign of peritectic transformation could be observed in the present work, it could not be totally excluded that alpha transformed fully in beta due to slow cooling. An assumption could be made that it was thought possible that nucleation of alpha was favoured with respect to beta. Competitive nucleation of phases from the liquid is known to be influenced by the cooling rate condition, as noted by Langsurd [3] who reported the shift of phase boundaries to a higher iron content and lower silicon with increased cooling rate. The shift may be due either to delayed nucleation or to a change in solidification path. Therefore, in order for the alloy to have a hexagonal alpha precipitates, the solidification path should diverge to lower silicon and higher iron content. However, this condition more likely is not going to happen due to the homogeneity of element within the liquid.

Further analysis for the investigation of Chinese script phase nucleation was performed with DTA. The experiment was conducted at cooling rate of 0.05°C/min and reheating (2°C/min) the sample prior to the beta phase precipitation. If hexagonal alpha did precipitate because of more favourable nucleation kinetics than for beta, its precipitation should occurr first. The thermal arrest detected between the liquidus and beta precipitation (Figure 6.1) could then be associated with the script phase. However, the DTA thermogram of reheating sample did not show any thermal arrest that could be associated with dissolution of an iron rich phase, as seen in Figure 6.11. The small deviation observed before the liquidus peak prior returning to baseline (black arrow) is related to DTA apparatus characteristic. This result showed that the script-beta phase might have appeared together with the plate-like beta phase.

Figure 6.11. DTA thermogram during of cooling down to a temperature higher than for beta precipitation and then reheating to melting temperature.

Metallographic evaluation showed that there are no plate-like beta precipitates in the regions where the script-beta precipitates are observed. SEM micrograph of deep etched sample showed the script phase in the interdendritic region is surrounded by eutectic silicon as seen in Figure 6.12.

Figure 6.12. Script phase (white arrow) shown attached to eutectic silicon in the interdendritic region after deep etch.

Metallographic observation indicated the presence of another phase near or within the script precipitates, in the form of small plates attached or small particles dispersed within the script precipitates as illustrated in Figure 6.13. EDS analysis showed this phase to be a cerium rich phase, as seen in Figure 6.6. Cerium is known to improve the mechanical properties

590 595 600 605 610 615 620 625

temperature (°C)

decreases the temperature for (Al) nucleation and might act as grain refiner, but no effect on iron-rich phases has been reported. If we exclude the script like precipitates have first formed as alpha, then this means they are a growth form of beta. Cerium should act on the growth, not necessarily on nucleation. Cerium precipitates thus indicate where this element was available.

Figure 6.13. Optical micrographs showing the presence of Ce-rich phase attached and dispersed in the script phase (red arrow).

The mechanism for the morphology changes of beta phase could be related to the cooling rate. In low cooling condition, Chinese script beta precipitates were seen in interdendritic region. As the cooling rate increases, the phase becomes smaller and more difficult to distinguish in the matrix.

Ce-rich phase precipitation can be seen by following the liquidus projection of Al corner of the Al-Si-Ce phase diagram [8] as illustrated in Figure 6.14. For an alloy with low level of cerium, the most possible Ce-rich intermetallic to appear during solidification is 2

(AlCeSi2) according to the tentative solidification path the eutectic line E2(L 2+ (Al) + Si, 573 °C) reaction. At low cooling rate Ce-rich phase has enough time to precipitate while as the cooling rate increases the Ce-rich precipitates became difficult to detect and may have not had time to appear and grow.

Figure 6.14. Liquidus projection of Al corner of Al-Si-Ce phase diagram [8].